• Journal of KIISE:Software and Applications
• /
• v.35 no.3
• /
• pp.179-188
• /
• 2008

An important issue in multiagent reinforcement learning is how an agent should team its optimal policy through trial-and-error interactions in a dynamic environment where there exist other agents able to influence its own performance. Most previous works for multiagent reinforcement teaming tend to apply single-agent reinforcement learning techniques without any extensions or are based upon some unrealistic assumptions even though they build and use explicit models of other agents. In this paper, basic concepts that constitute the common foundation of multiagent reinforcement learning techniques are first formulated, and then, based on these concepts, previous works are compared in terms of characteristics and limitations. After that, a policy model of the opponent agent and a new multiagent reinforcement learning method using this model are introduced. Unlike previous works, the proposed multiagent reinforcement learning method utilize a policy model instead of the Q function model of the opponent agent. Moreover, this learning method can improve learning efficiency by using a simpler one than other richer but time-consuming policy models such as Finite State Machines(FSM) and Markov chains. In this paper. the Cat and Mouse game is introduced as an adversarial multiagent environment. And effectiveness of the proposed multiagent reinforcement learning method is analyzed through experiments using this game as testbed.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.8 no.4
• /
• pp.123-133
• /
• 2008

This study evaluated the runoff data collected at 12 stream gauge stations of the Chungjoo dam basin using the IHACRES model. Especially, the geomorphology-related parameters of the IHACRES model could be quantified base on the regionalization technique, which have also been applied many stream gauge stations of the Chungjoo dam basin. Summarizing the results is as follows. (1) The climate-related parameters of the IHACRES model c, $\tau_w{^0}$, and f are found to be estimated and used uniformly over the basin. (2) The geomorphology-related parameters of the IHACRES model $t_q,\;t_s,\;and\;v_s$ are found to be estimated by considering the geomorphological parameters like the basin area, channel length, channel slope, basin slope through the regionalization based on the regression analysis. (3) Using the climate-related parameters applied uniformly over the basin and the geomorphology-related parameters estimated based on the regionalization procedure for each stream gauge station, a total of 12 stream gauge stations have been evaluated with their stream flow measurements. As results, the Sanganmi and Youngwal 1 stream gauge stations have been found to make high quality flow data, but Youngwal, Baekokpo, and Panwoon stations low quality flow data. On the whole, 12 stream gauge stations considered show large differences with their data quality, so a plan for securing more consistent data quality should be prepared imminently.

• Journal of Life Science
• /
• v.27 no.4
• /
• pp.435-441
• /
• 2017

Perchlorate ($ClO_4^-$) is an emerging contaminant detected in soil, groundwater, and surface water. Previous study revealed bacterial community in the enrichment culture tdegraded perchlorate using elemental sulfur as an electron donor. Quantitative and qualitative molecular methods were employed in this study to investigate archaeal community in the enrichment culture. Real-time qPCR showed that archaeal 16S rRNA gene copy number in the culture was about 1.5% of bacterial 16S rRNA gene copy number. This suggested that less archaea were adapted to the environment of the enrichment culture and bacteria were dominant. DGGE banding pattern revealed that archaeal community profile of the enrichment culture was different from that of the activated sludge used as an inoculum for the enrichment culture. The most dominant DGGE band of the enrichment culture was affiliated with Methanococci. Further research is necessary to investigate metabolic role of the dominant archaeal population to better understand microbial community in the perchlorate-reducing enrichment culture.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.6
• /
• pp.555-565
• /
• 2021

The goal of this study is to improve flow and noise performances of existing backward-curved blade centrifugal fan system used for circulating cold air in a refrigerator freezer by optimally designing airfoil shape. The unique characteristics of the system is to drive cold airflow with two volute tongues in combination with duct system in a back side of a refrigerator without scroll housing generally used in a typical centrifugal fan system. First, flow and noise performances of existing fan system were evaluated experimentally. A P-Q curve was obtained using a fan performance tester in the flow experiment, and noise spectrum was measured in an anechoic chamber in the noise experiment. Then, flow characteristics were numerically analyzed by solving the three-dimensional unsteady Navier-Stokes equations and noise analysis was performed by solving the Ffowcs Williams and Hawkins equation with input from the flow simulation results. The validity of numerical results was confirmed by comparing them with the measured ones. Based on the verified numerical method, blade inlet and outlet angles were optimized for maximum flow rate using the two-factor central composite design of the response surface method. Finally, the flow and noise performances of a prototype manufactured with the optimum design were experimentally evaluated, which showed the improvement in flow and noise performance.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.697-703
• /
• 2022

Due to the recent economic downturn caused by Covid-19 and the unstable international situation, many investors are choosing the derivatives market as a means of investment. However, the derivatives market has a greater risk than the stock market, and research on the market of market participants is insufficient. Recently, with the development of artificial intelligence, machine learning has been widely used in the derivatives market. In this paper, reinforcement learning, one of the machine learning techniques, is applied to analyze the scalping technique that trades futures in minutes. The data set consists of 21 attributes using the closing price, moving average line, and Bollinger band indicators of 1 minute and 3 minute data for 6 months by selecting 4 products among futures products traded at trading firm. In the experiment, DNN artificial neural network model and three reinforcement learning algorithms, namely, DQN (Deep Q-Network), A2C (Advantage Actor Critic), and A3C (Asynchronous A2C) were used, and they were trained and verified through learning data set and test data set. For scalping, the agent chooses one of the actions of buying and selling, and the ratio of the portfolio value according to the action result is rewarded. Experiment results show that the energy sector products such as Heating Oil and Crude Oil yield relatively high cumulative returns compared to the index sector products such as Mini Russell 2000 and Hang Seng Index.

• Journal of Korea Port Economic Association
• /
• v.34 no.3
• /
• pp.93-112
• /
• 2018

The purpose of this study is to analyze the flow of e-commerce freight transported by maritime transportation for China and to identify the characteristics of cargo by region/item for finding the ways to promote e-commerce export to China. Thus, this study analyzed the e-commerce export and import data on cargo moved via maritime transportation between Korea and China from 2015 to 1Q18, using Origin-Destination(OD) analysis and visualization techniques. The results indicated that the largest number of Chinese e-commerce cargoes were imported at Incheon Port, which has a clearance facility for e-commerce cargo. In the case of Pyeongtaek Port, e-commerce cargo imported from China has transported to Incheon Customs again, causing the inefficiency through the customs clearance process. Unlike the case of e-commerce imports where the final destination is distributed nationwide, e-commerce products exported to China through maritime transportation were found to be mainly confined to Seoul and Gyeonggi provinces, where freight forwarding companies and forwarders are concentrated. In addition, unlike e-commerce import cargoes, e-commerce items exported through maritime transportation were mainly confined to clothing and cosmetics, and export volume was also less than imports. This study provides some possible strategies to increase the volume of freight and to attract export products as follows: i) to diversify products exported to China through e-commerce transshipment, ii) to diversify export items by building the cold chain in e-commerce transport with China.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.15 no.11
• /
• pp.1043-1050
• /
• 2004

In this paper, Phase noise is analyzed and a novel approach of the nonlinear approximation including second order term of phase noise is presented to analyze and quantize system performance. As results, in QPSK-OFDM system, when PLL loop bandwidth is 5.0 Hz, 1.0 kHB, 0.5 kHz respectively, there are about 0.6 dB, 1.0 dB, 1.7 dB SNR penalties at BER=10$\^$-4/ compared with system without phase noise in AWGN channel. In 16QAM modulation, there are about 1.9 dB, 3.2 dB, 6.7 dB SNR penalties at BER=10$\^$-4/ respectively. At QPSK-OFDM system, comparing the previous linear approximation method with our proposed nonlinear approximation method, there is similar BER performance at phase noise variance lower than 0.02, but certain difference occurs as variance increases more than 0.02. Furthermore, analytical BER results closely match with simulation results in the OFDM system employing QPSK and 16qAM modulation. And, BER performance of QPSK-OFDM system is considerably degraded because of the BER error floor if the phase noise variance becomes larger than 0.03.

• Geotechnical Engineering
• /
• v.13 no.6
• /
• pp.95-106
• /
• 1997

A pressuremeter is an expandable tube which is placed in the soil, and then expanded under controlled condition against the soil. From this test a pressure expansion curve of the soil can be obtained. However soil disturbance during the test has significant influence on the results of tests. A general governing equation for pressuremeter test can be theoretically derived on the basis of the hyperbolic soil model and the cavity expansion theory. The curve fitting technique was used to establish the pressure-strain curve without disturbance of soil during testing. This interpretation makes use of both the loading and unloading portions of the test. An interpretation methodology is described and illustrated with pressuremeter test data carried out in the weathered granitic soil to estimate initial shear modulus. Standard penetration test is a very common site investigation technique in Korea. Therefore the blow counts of standard penetration test are discussed by comparing them with initial shear modulus.

• Journal of IKEEE
• /
• v.23 no.3
• /
• pp.1112-1115
• /
• 2019

This paper proposes a novel low-profile omni-directional microstrip antenna to mount on the deployable wing of the observation munition. The proposed antenna is designed on seven hexagonal resonators in a quasi-circular array to achieve a monopolar radiation pattern with a thin substrate. By employing the mesh ground structures, the resonant frequencies and impedance bandwidths of the proposed antenna is investigated. To verify the feasibility of the mesh ground structure, the thin ground wire width is investigated theoretically for improving the 3-dB fractional bandwidth, realized antenna gain and quality-factor. The proposed antenna demonstrates a good monopolar radiation in good agreement with the simulation results. The implemented prototype shows the measured bandwidth of 326 MHz with respect to 5.65% centered at 5.77 GHz and realized gain of 5.49 dBi at 5.84 GHz.

• Journal of Digital Convergence
• /
• v.20 no.3
• /
• pp.201-209
• /
• 2022

In order for the military to review the introduction of virtual reality technology into various education and training systems and fully utilize it, it is important to reflect the characteristics of the technology and education system and to accurately identify and selectively apply the characteristics of commercialization. In this study, the evaluation criteria were selected through the Analitic Hierarchy Process (AHP) method for factors to be considered when commercializing a virtual reality-based education and training system, and the priorities of the projects were determined. Based on previous studies, an initial AHP model was constructed and the relative importance of six factors, including reality, was analyzed as the level 1 evaluation criteria. Next, for Level 2, each evaluation criterion was evaluated to confirm the importance of each of the 11 tasks in the six evaluation criteria, and priorities were selected for each task. As a result of the analysis, level 1 showed that reality and ripple had higher importance than other factors. As a result of evaluating the final relative importance, the priority was shown in the order of ① flight training, ② disaster training, ③ shooting Training, and ④ driving a vehicle. Based on the relative priorities determined in Levels 1 and 2 of the model presented in this study, the importance of each project necessary for final decision-making of the research priorities for the defense virtual reality project was presented. It is expected that this study can be used as a reference material for prioritizing the commercialization of education and training systems in the defense sector.